An Efficient Approach of Video Based Face Recognition and People Counting System

Abstract

In recent years, face detection is widely used in various fields, such as face recognition, image focusing, and surveillance systems. It makes another way in the biometrics field. A Support Vector Machine based multi-view face detection and recognition framework is described in this paper. This study proposes a real-time face detection system based on Support Vector classifier using CNN. The detection system divided into three main parts, first detect face, two CNN feature extractor that generates 128-d facial embedding, three train a support vector machine (SVM) on top of the embedding, four recognize faces in images and video streams. We’ll be applied deep learning in two key steps, first to apply face detection, which detects the presence and location of a face in an image, but does not identify it, second to extract the 128-d feature vectors (called “embedding”) that quantify each face in an image. These face embedding will be sufficiently different such that we can train a “standard” machine learning classifier SVM on top of the face embedding. In this paper, face recognition system to implement in CNN. The experimental results show that the accuracy rate is higher than 86.67% in face detection, which implies the proposed real-time detection system is indeed effective and efficient. We present experimental consequences of our implementation of SVM, and demonstrate the possibility of our methodology on face identification, issue that includes a data set of 10,000 data points. Detailed experimental results are presented in this report including tuning the parameters of the face detectors, performance evaluation, and applications to video based face detection and frontal-view face recognition.